#
✕ nick:maz
#asahi-dev · 2024-01-13 12:02
mazMinor Central Scrutinizer update pushed out (fixed LCSC references, nothing major), gerber and bom files updated.
#asahi-dev · 2024-01-06 17:10
mazjanneg: perfect thank you very much. I think I can safely wait for M4! :D
#asahi-dev · 2024-01-06 16:56
mazjanneg: awesome thanks. So far very close to M2. Anything in MMFR2/ISAR2?
#asahi-dev · 2024-01-06 14:13
mazjanneg: when you get a chance, could you dump the M3 CPU idregs somewhere?
#asahi-dev · 2024-01-04 14:56
mazah, right. as you can tell, FP isn't really my thing! :D
#asahi-dev · 2024-01-04 14:48
mazif memory serves well, the rounding modes are different between x86 and ARMv8, unless the implementation has FEAT_AFP (and SW picks the correct mode).
#asahi-dev · 2024-01-04 14:21
mazthe question about semantics still applies though. not sure the BFloat16 format is the same as the "half-float" leio is talking about.
#asahi-dev · 2024-01-04 14:18
mazj`ey: M2 has FEAT_BF16, but M1 doesn't.
#asahi-dev · 2023-11-30 17:07
mazyou did a brilliant job, congrats!
#asahi-dev · 2023-11-30 12:20
mazyeah, same here. fortunately, I managed to sell NV as being relevant to $work. it is quite a stretch of one's imagination, but hey...
#asahi-dev · 2023-11-30 12:17
mazsven: and you'll spend another 3 years reverse-engineering it! ;-)
#asahi-dev · 2023-11-30 12:15
maz(hopefully no major new feature so that I don't have to start saving money... :D)
#asahi-dev · 2023-11-30 12:15
mazany chance you could share the idregs?
#asahi-dev · 2023-11-30 12:14
mazDannyB: amazing!
#asahi-dev · 2023-11-30 07:24
mazyou just have to sort out the interrupt controller now...
#asahi-dev · 2023-11-30 07:24
mazin any case, congrats, and good that you got to the bottom of it!
#asahi-dev · 2023-11-30 07:23
mazwell, that's just a lot of memory! :D
#asahi-dev · 2023-11-29 17:51
mazplays Judas Priest's Breaking The Law...
#asahi-dev · 2023-11-29 07:24
mazDannyB: awesome, thanks for letting me know!
#asahi-dev · 2023-11-28 16:16
maz11.0 or similar.
#asahi-dev · 2023-11-28 16:15
mazI'm still on whatever the thing came with.
#asahi-dev · 2023-11-28 16:14
mazmaybe the FW on my M1 is a bit too old (never updated).
#asahi-dev · 2023-11-28 16:03
mazhuh, I never realised: generating a reset command from a CS on a M2 that is powered-off brings it back to life (tested on M2 Pro mini and M2 MBA). doesn't work on my M1 mini.
#asahi-dev · 2023-11-28 14:35
mazdumping the id registers would settle it.
#asahi-dev · 2023-11-28 14:21
mazinteresting!
#asahi-dev · 2023-11-28 14:14
mazif they have stuffed non-architectural stuff in TCR_EL2, that's going to be fun. not.
#asahi-dev · 2023-11-27 10:19
mazsven: TB DART merged. good stuff!
#asahi-dev · 2023-11-24 10:00
mazextremely impressive indeed. though I'm in the same situation as j`ey on this front!
#asahi-dev · 2023-11-21 12:09
mazmarcan: cheers.
#asahi-dev · 2023-11-21 11:59
mazon the positive side, I've booted Juno for the first time in two years. didn't miss it.
#asahi-dev · 2023-11-21 11:57
mazmarcan: of course. that's definitely bug report material. linux-perf-users@vger.kernel.org and LKML are the advertised lists.
#asahi-dev · 2023-11-21 11:50
mazso yes, anything BL is dead.
#asahi-dev · 2023-11-21 11:50
mazah, even Juno is broken. brilliant.
#asahi-dev · 2023-11-21 11:46
mazI sort of suspected this... the author had some pretty odd ideas on the list...
#asahi-dev · 2023-11-21 11:40
mazI'm pretty sure the PERF_TYPE_HARDWARE changes are related to this breakage.
#asahi-dev · 2023-11-21 11:35
mazAh, fun. just built from -rc2, and it indeed outputs crap. it's probably broken for any flavour of non-x86 asymmetric system...
#asahi-dev · 2023-11-21 11:26
mazwhatever Debian ships.
#asahi-dev · 2023-11-21 11:25
mazsame*
#asahi-dev · 2023-11-21 11:25
mazme version on all 4 boxes.
#asahi-dev · 2023-11-21 11:24
mazperf version 6.1.55 on M2 MBA
#asahi-dev · 2023-11-21 11:24
mazM2 MBA works fine with 6.6-rc6.
#asahi-dev · 2023-11-21 11:23
mazlet me try
#asahi-dev · 2023-11-21 11:22
mazbasic m2?
#asahi-dev · 2023-11-21 11:22
mazwhich one is t8112?
#asahi-dev · 2023-11-21 10:59
mazwent back to 6.5 on M1 mini, and it works fine.
#asahi-dev · 2023-11-21 10:58
mazis it*
#asahi-dev · 2023-11-21 10:58
mazis a universal problem on all machines, or something specific to t600x?
#asahi-dev · 2023-11-21 10:54
mazhow? they are per-CPU.
#asahi-dev · 2023-11-21 10:54
mazactually, my reasoning doesn't hold. there are no interrupts firing for such a test.
#asahi-dev · 2023-11-21 10:52
mazah.

plus ›